4 to 2
Browse files
README.md
CHANGED
@@ -13,9 +13,9 @@ tags:
|
|
13 |
- macadeliccc/Samantha-Qwen2-1.5B
|
14 |
---
|
15 |
|
16 |
-
# Qwen2-
|
17 |
|
18 |
-
Qwen2-
|
19 |
* [cognitivecomputations/dolphin-2.9.3-qwen2-1.5b](https://huggingface.co/cognitivecomputations/dolphin-2.9.3-qwen2-1.5b)
|
20 |
* [macadeliccc/Samantha-Qwen2-1.5B](https://huggingface.co/macadeliccc/Samantha-Qwen2-1.5B)
|
21 |
|
@@ -68,7 +68,7 @@ from transformers import AutoTokenizer
|
|
68 |
import transformers
|
69 |
import torch
|
70 |
|
71 |
-
model = "djuna/Qwen2-
|
72 |
|
73 |
tokenizer = AutoTokenizer.from_pretrained(model)
|
74 |
pipeline = transformers.pipeline(
|
|
|
13 |
- macadeliccc/Samantha-Qwen2-1.5B
|
14 |
---
|
15 |
|
16 |
+
# Qwen2-2x1.5B
|
17 |
|
18 |
+
Qwen2-2x1.5B is a Mixture of Experts (MoE) made with the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
|
19 |
* [cognitivecomputations/dolphin-2.9.3-qwen2-1.5b](https://huggingface.co/cognitivecomputations/dolphin-2.9.3-qwen2-1.5b)
|
20 |
* [macadeliccc/Samantha-Qwen2-1.5B](https://huggingface.co/macadeliccc/Samantha-Qwen2-1.5B)
|
21 |
|
|
|
68 |
import transformers
|
69 |
import torch
|
70 |
|
71 |
+
model = "djuna/Qwen2-2x1.5B"
|
72 |
|
73 |
tokenizer = AutoTokenizer.from_pretrained(model)
|
74 |
pipeline = transformers.pipeline(
|